pyppe.clj

(ns lightusers.core
  (:require [clojure.data.json :as json]
            [clojure.pprint :refer [pprint]]))

(def json-pprint (comp pprint json/write-str))
(def input (json/read-str (slurp "example.json")))


(defn extract
  ([tweet paths] (extract tweet paths []))
  ([tweet paths users]
   (if (empty? paths)
     {:light-users-tweet tweet :extracted-users users}
     (let [path (first paths) tail (next paths)]
       (if-let [user (get-in tweet path)]
         (extract
          (update-in
           tweet path select-keys ["name" "id_str"])
          tail (conj users user))
         (extract tweet tail users))))))


(def paths [["user"]
            ["retweeted_status" "user"]
            ["retweeted_status" "quoted_status" "user"]
            ["quoted_status" "user"]])

(let [output (extract input paths)]
  (println "EXTRACTED USERS:")
  (map json-pprint (:extracted-users output))
  (println "WHOLE TWEET WITH LITE USERS:")
  (json-pprint (:light-users-tweet output)))